XLS-R-300M Open-Source Speech Recognition Model - Free Deployment for Precise Dutch Speech Recognition

Xlsr300m Cv 7.0 Nl Lm

Developed by Iskaj

XLS-R-300M is an automatic speech recognition (ASR) model specifically optimized for Dutch, trained on the Common Voice 8 Dutch dataset.

Downloads 23

Release Time : 3/2/2022

Model Overview

This model is primarily used for automatic speech recognition tasks in Dutch, capable of converting Dutch speech into text.

Dutch optimization

Specifically optimized for Dutch, providing better speech recognition performance.

Multi-dataset evaluation

Performance evaluated on Common Voice 8 and robust speech event datasets.

Medium-sized model

300M parameter size, balancing performance and computational resource requirements.

Dutch speech recognition

Speech-to-text

Speech recognition

Dutch speech transcription

Convert Dutch speech content into text format.

On the Common Voice 8 Dutch test set, WER is 32 and CER is 17.

Robust speech recognition

Perform Dutch speech recognition in complex environments.

On the robust speech event development data, WER is 37.44, and on the test data, WER is 38.74.

Property	Details
Model Name	XLS-R-300M - Dutch
Task	Automatic Speech Recognition
Datasets	Common Voice 8 NL (`mozilla-foundation/common_voice_8_0`), Robust Speech Event - Dev Data (`speech-recognition-community-v2/dev_data`), Robust Speech Event - Test Data (`speech-recognition-community-v2/eval_data`)
Metrics
- Test WER on Common Voice 8 NL	32
- Test CER on Common Voice 8 NL	17
- Test WER on Robust Speech Event - Dev Data	37.44
- Test WER on Robust Speech Event - Test Data	38.74

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base